Tyrosine kinase inhibitors. 6. Structure-activity relationships among N- and 3-substituted 2,2'-diselenobis(1H-indoles) for inhibition of protein tyrosine kinases and comparative in vitro and in vivo studies against selected sulfur congeners.

Abstract

A small series of 2,2'-diselenobis(1H-indoles) was synthesized as redox-modified congeners of our earlier reported 2,2'-dithiobis(1H-indole) series. Utilizing chemistry similar to that developed earlier for the disulfur series, compounds were made from 2-halogeno-3-indolecarboxylic acid precursors bearing various polar functionality at the C-3 position and small alkyl substituents at the N-1 position of the indole nucleus. Additional compounds were derived from (R)- or (S)-tryptophan via a novel application of diselenium dichloride as an electrophilic source of diselenium, and a much improved process to a 2,2'-dithiobis(1H-indole) congener was developed utilizing disulfur dichloride as a source of disulfur. Against isolated epidermal growth factor receptor (EGFr), platelet-derived growth factor receptor (PDGFr), and v-src tyrosine kinases, compounds in this series displayed broad inhibitory activity with IC50 = 0.9 to > 100 microM vs EGFr, 3.4 to > 50 microM vs PDGFr, and 0.4-6.7 microM vs v-src. In general, compounds derived from tryptophan displayed the greatest potency against EGFr and those from 2-halogeno-3-indolecarboxylic acids greater potency against PDGFr and v-src. Enzyme kinetics studies showed that both classes of compounds display primarily noncompetitive inhibition with respect to either ATP or peptide substrate. The sulfhydryl reducing agent dithiothreitol (DTT) caused a general decrease in inhibition of the EGFr and v-src tyrosine kinases by both the diselenium and disulfur series with the reversal of enzyme inhibition occurring less readily within the diselenium series. In whole cell studies, compounds of this class were growth inhibitory against Swiss 3T3 mouse fibroblasts with IC50 values from 0.5 to 19.5 microM, and the observed SAR was different from that of the 2,2'-dithiobis(1H-indoles). A comparative study in the same cell line on the effects of the 2,2'-diselenobis(1H-indole) derived from (R)-tryptophan vs its disulfur congener on growth factor mediated tyrosine phosphorylation showed that this compound significantly inhibited EGFr and PDGFr (in response to its ligand) autophosphorylation with complete suppression at 25 and 5 microM, respectively. Tyrosine phosphorylation of an 85 kDa protein typically phosphorylated in response to bFGF was also exquisitely sensitive to this compound, and it displayed inhibitory effects on DNA, RNA, and protein synthesis at submicromolar concentrations. The disulfur congener exhibited a qualitatively similar pattern; however, its potency was 10-fold less. This same diselenium/disulfur pair was evaluated in vivo against the B16 melanoma, colon carcinoma 26, and M5076 sarcoma murine tumors, and the A431 epidermoid, and C6 glioma human tumor xenografts. At maximum tolerated doses (1.8 and 5.0 mg/kg/injection, respectively), neither the diselenium nor disulfur congener was effective against the C6 glioma when administered intraperitoneally on a d1-9 schedule. Studies were also carried out against the A431 epidermoid xenograft to evaluate the same pair of compounds via continuous subcutaneous infusion from Alzet miniosmotic pumps. The maximum dose that could be administered daily was limited by compound solubility. Neither compound produced an antitumor effect in a 7-day continuous infusion study. In the 27-day study, the disulfur compound was inactive whereas the diselenium compound produced a 10.8-day growth delay without appreciable treatment related weight loss. The in vitro and in vivo findings offer a mechanistic rationale as to why the 2,2'-diselenobis(1H-indoles) are more potent inhibitors than their disulfur congeners.